Tuhin Mallick

Building AutoML solutions for forecasting - Responsible for developing AI-based algorithms for AutoML service for time series forecasting - Building sklearn-like for end-to-end time series forecasting as web service - Technologies include: Python as programming language, Azure devops, Argo workflow management, FastAPI - Algorithms include: -econometric time series model -classic ML & deep learning for time series forecasting -Facebook Prophet-feature selection, feature engineering & model interpretation

I am highly focused on artificial intelligence and computer vision and love to contribute to the Open-source project I have recently concluded an open-source project in collaboration with Grau Data. The project was in line with Industrial Agile and XP programming practices where we worked on Synthetic File System to be used as a pre-processing pipeline for any artificial intelligence software. -- Synthetic File System ( https://github.com/amosproj/amos2021ss03-synthetic-file-system)

Conception and implementation of an automation solution for flexible screw separation using Robot Operating System and Artificial intelligence. Deep Object Pose Estimation (DOPE) is being used to determine the pose while YOLO is being used to minimize the area of application of the pose estimator. The 3D object models are created using the Blender tool, which is used for generating synthetic training datasets with the help of Unreal Engine (UE4) and NVidia Deep Learning Data Synthesizer (NDDS) software.

The OnHW Dataset: Online Handwriting Recognition from IMU-Enhanced Ballpoint Pens with Machine Learning. The novel OnHW-chars dataset allows for the evaluations of uppercase, lowercase and combined classification tasks, on both writer-dependent (WD) and writer-independent (WI) classes and we show that properly tuned machine learning pipelines as well as deep learning classifiers (such as CNNs, LSTMs, and BiLSTMs) yield accuracies up to 90 % for the WD task and 83 % for the WI task for uppercase characters.

During my tenure I have been working on a web application development project for the National Insurance Company and overseeing the construction of UI using jQuery, CSS and bootstrap responsive web design. I am using the Eclipse Rich client platform (RCP) for development and reuse of components.

Relevant Coursework : Computer Vision, Artificial intelligence, high performance computing, quantum computing Honors: Erasmus+ scholarship (2022/23)

Coursework includes Lectures and Projects on : 1. Pattern recognition 2. Deep Learning 3. Reinforcement learning 4. Optimization Honors : DAAD PROMOS scholarship and FAU Reisekostenstipendien

Robotics - Geek United Pravasana - Photography Club Rotaract Club of Hitk College Table tennis team College Atheletic Team